Estimation on Scale Error of SSC Retrieval Model Based on Scale Expansion Method

Abstract

Although the variance of suspended sediment concentration (SSC) distribution is an important parameter for estimating scale error of SSC retrieval results, in the past it has not been possible to use this parameter because the requisite comprehensive oceanic optical data sets have not been available. In this study, the approaches of scale expansion, correlation analysis, and regression analysis were used to reveal the relationship between the variance of SSC distribution and the SSC of central pixel, average SSC, and pixel size in a moving window. According to the study results, it was found that the variance of SSC distribution was highly correlated to the SSC of central pixel and average SSC in a given scale window, while less correlated to pixel size. Additionally, the relationship model between the variance of SSC distribution and average SSC in a given scale window was used to correct the scale errors in retrieval results combined with a scale error model. The results showed that the greatest scale error of SSC retrieval results was up to 30%.